How to use mispeech/dashengtokenizer with Transformers:
# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("mispeech/dashengtokenizer", trust_remote_code=True, dtype="auto")